Comparison of PI and fuzzy PI controls of synchronous buck dc converters

From the day when the electricity was invented to the day, numerous electrical devices have been invented. Many of these devices cannot be directly operated by the energy in the electricity grid. Because, while the electric network is in an alternating current waveform, many of the electric devices we use work with the direct current, or the voltage value on the network and the working voltage value of the device may be different. This leads to the necessity of conversion of electric energy from one wave form to another or from a voltage level to another voltage level. In this case, the power electronics step in and makie it possible to convert between alternating and direct current forms of electric energy and to set different levels in the same waveform. Power electronics is a sub-discipline of electrical engineering, which is concerned with circuits that convert from a waveform to other waveform or adjust the voltage and/or current level in many systems powered by electricity, from household appliances, electric motor drives, renewable energy sources to direct current high voltage transmission. Power electronics is basically based on semiconductor switching. With the development of semiconductor technology, the power, switching speeds and efficiencies that power electronics componenets can carry are increasing considerably. In addition, the development of microprocessor and microcontroller technologies has had a great impact on the development of power electronic control techniques. Thanks to this control power, power electronic circuits are used more effectively and more efficiently. Power electronic circuits are basically four different types of converters. These are from alternating current to alternating current, from direct current to alternating current, from alternating current to direct current, and from direct current to direct current. In this study, a synchronous buck direct current converter, one of the buck direct current converters, is dealt with. DA converters are widely used in electric automobiles, cranes and traction motors of conveyor belts for their control. DA converters are also one of the most important elements of energy conversion in the field of developing renewable energy. In addition, DA converter applications have become an indispensable part of consumer electronics (appliances, small appliances, televisions, mobile phone chargers). Freewheeling diodes are used in conventional buck converters cause significant power losses in high-power applications and also have a dimensional disadvantage in design. MOSFET switching design instead of using freewheeling dioede provide more efficiency. So the design is called synchronous buck converter using both on high side and low side MOSFETs. Direct current synchronous converters operate continuously and are suitable for two-way operation. In synchronous buck converters there are two operating state. High side MOSFET is on, low is off and high side MOSFET is off and low side is on state. This states take place periodically and it determine the duty ratio of synchronous buck converter and it will be used as control variable in control of the circuit. The capacitor and inductor values to be used in the direct current converter are derived from the Kirchhoff voltage equations written for two operating states. Switching frequency and duty ratio are parameters that should be determined in before when selecting capacitor and inductor. The connection of the selected inductors and capacitors is in the low-pass filter structure, which determines the output ripple. Control is the other important issue in the circuit design. Designed with a closed-loop control method, it is possible to operate the circuit in a certain operating range despite the disturbing effects. As in many applications, PI control can provide a very effective, easy and inexpensive control for DC converters. With P control, the P control coefficient Kp is multiplied by the error to determine how much the output will change. Since the error signal is very large initially, the proportional controller responds quickly and this effect diminishes as the error signal shrinks. In proportional control, since the error signal is multiplied by the value Kp, the control signal can never be zero, and a steady state error occurs. I control coefficient KI effect produces a control signal that increases with positive errors and decreases with negative errors. In this control system, where the sum of positive and negative error signals are continuously multiplied by an integral constant, the continuous sum of the error signals will decrease after a while and eventually the steady state error will be zero. PI control consists of a combination of P and I control and it is to approach the reference value rapidly by the effect of the proportional proportional component and to get rid of the permanent state error by the effect of the integral component. One of the most commonly used methods for determining the coefficient is the Ziegler-Nicholes method because the coefficients can be easily obtained. KP and KI values were determined by Ziegler Nichols method with unit step response. In addition, it can be considered that the adaptive method which will change the KP and KI coefficients instead of fixed Kp and KI coefficients in the classical PI control, can be provide more effective control. When the previous studies were examined, Fuzzy logic based PI control showed more successful results than PI control. Thus, the motivation for working this thesis is provided. To design a fuzzy logic based PI control circuit, basic concepts and applications of fuzzy control such as fuzzy set, membership function, fuzzy inference, and defuzzification are needed. In this study, the fuzzy logic controller uses error and change of error as fuzzy input variables. For these input values, appropriate input membership functions have been determined either by taking advantage of expert knowledge or taking into account similar studies that have been done previously. Then the appropriate rule table is designed for the system. The input values are written in the respective membership functions for each rule and the fuzzy outputs are obtained as a result. The fuzzy outputs are also written to the designated output membership functions to determine the actual output values. Finally, the desired output value is obtained as a result of the defuzzification stage. Thus, a fuzzy logic controller was designed to determine the Kp and KI coefficients and a fuzzy logic based PI structure was created. In this thesis, it is aimed to design a synchronous buck converter and to compare and contrast the results of classical PI control and Fuzzy PI control as a method for adjusting Kp and KI coefficients according to system response. The mathematical model of the constant state of the system was obtained for the selection of circuit components. In order to examine the stability of the cycle and to determine the KP and KI parameters, the transient state mathematical model and transfer function were obtained. Simulation studies were performed on MATLAB / Simulink to examine the results of PI control and fuzzy PI control. As a result of the simulated experiments, it was observed that fuzzy PI control gives better results in parameters such as rise time, settling time compared to PI control. After simulation work, circuit design has come to the stage. For the design of the circuit, the switching frequency and the operating ratio, the current ripple and the voltage ripple ratio are determined, and the capacitance and inductance values obtained from the calculations made on the axis of these parameters are found and the capacitor and the inductor are selected. The switching elements to be used in the circuit are selected by examining the catalogs considering some parameters like circuit output current, voltage values, switching frequency, on time and off time intervals, etc. A driver circuit is needed to drive the switching components and the bootstrap method, which is one of the most suitable driving methods, has been used for this study. In order to drive by bootstrap method, the necessary driver circuit is designed by selecting an appropriate Integrated Circuit element from the catalogs. After the design of the circuit has been realized, passed to the design phase of the controller to provide control of the circuit. The Atmel processor-based Arduino MEGA is used for control, and the algorithms for PI and fuzzy PI are coded in the Arduino compiler. The designed PI and Fuzzy PI controllers are integrated in the designed circuit and the circuit is operated. The resulting laboratory studies were compared for PI and Fuzzy PI control and the results were parallel to the simulation studies. As a result of laboratory studies, it was observed that fuzzy PI control gave better results in parameters such as rise time, settling time compared to PI control.